ABSTRACT
We have seen methods for solving linear systems or for finding eigenvalues. In practical applications, these methods may not be very efficient. For example, it is not numerically feasible to compute the zeroes of a 1000th-degree polynomial, as one would have to do for finding the eigenvalues of a 1000 × 1000 matrix. For such cases,
zero entries. An example of such a matrix is illustrated in Figure 16.1. Efficient handling of matrices such as this one is another motivation for developing special numerical methods.
